Changes in version 1.1.8                        

Highlights

  - Paper published in NAR Genomics and Bioinformatics (2025), Editor’s
    Choice (DOI: 10.1093/nargab/lqaf131).
  - Support for raw RNA-seq counts (requires gene lengths).
  - iBreastSubtypeR refresh: cleaner UX, smarter AUTO guidance,
    consistent exports.
  - Refined, data-driven thresholds for ER/HER2 skew detection in AUTO.
  - Broader input-validation across the subtyping pipeline.

Bug Fixes

  - AUTO: fixed ER/HER2 strata messaging for ER+/HER2− and ER−/HER2−
    cohorts. The spurious message “ssBC.v2 for samples: ERnegHER2neg” no
    longer appears in ER+/HER2− cohorts.
  - AUTO: in pure HER2+ cohorts, ssBC.v2 is now preferred as intended;
    ssBC is not listed or subset-indexed there.
  - AUTO: ssBC / ssBC.v2 sample-subsetting and messages now occur only
    when that method is actually selected.
  - PAM50 variants no longer error with calibration = "None" or
    "External".
  - ssBC variants handle datasets with <50 PAM50 genes more robustly.
  - Fixed data.frame issue (“check.names matched by multiple arguments”)
    via safe builders.
  - Removed duplicate “Subtype == BS_*class” columns in downloads.
  - Safer ROR merges on PatientID with clearer notifications.
  - Eliminated jsonlite named-vector warning in plotting.

Changes

  - Documentation: clarified guidance for subtype-specific cohorts:
      - ER/HER2-defined cohorts (ER+/HER2−, ER−/HER2−, ER+/HER2+,
        ER−/HER2+): NC-based → ssBC.v2 only; plus SSP-based (AIMS,
        sspbc).
      - ER-only (ER+ or ER−) and TNBC: NC-based → ssBC and/or ssBC.v2
        (subject to minimum sizes); plus SSP-based.
  - AUTO clarifications (simulation-based defaults):
      - ER balance gate: lower_ratio = 0.39, upper_ratio = 0.69.
      - Minimum sizes: ER+ total = 15, ER− total = 18, TN total = 18.
      - Subgroup gates (rounded): ER+ subgroups (HER2+ / HER2−) use
        round(15 / 2); ER− subgroups (HER2+ / HER2−) use round(18 / 2).
      - Notes: these thresholds gate method eligibility; they do not
        force a consensus call. Details in the vignette (“AUTO logic”).

Shiny (iBreastSubtypeR)

  - New hero card + method chips (NC, SSP, ROR, AUTO) and “Why AUTO?”
    explainer modal.
  - AIMS: 4-class toggle disabled (AIMS is 5-class only).
  - AUTO/AIMS/SSPBC: Full-metrics export selector hidden.
  - Clear per-method help + PAM50 calibration notes.
  - Small UX polish: file dialogs + Mapping preserve scroll position.

Exports

  - Two modes: Calls only and Full metrics (incl. ROR when available).
  - Standard column names: Call_5class / Call_4class (internal BS /
    BS.Subtype mapped).
  - AUTO: Calls = selected-k table (+ entropy); Full = selected-k merged
    with other-k (suffix _4class / _5class).
  - Filenames encode method, class, and mode (e.g.,
    results-PAM50-5class-full.txt).

Behavior & Validation

  - ROR only for NC methods when TSIZE/NODE exist and are numeric (auto
    coercion + warnings).
  - Stricter checks for ssBC subgroups (ER/HER2/TN presence and coding).
  - AUTO preflight warns on missing or mis-coded ER/HER2.

Core Features (unchanged)

  - Unified NC (PAM50, cIHC, PCAPAM50, ssBC) and SSP (AIMS, SSPBC)
    subtyping under one API.
  - BS_Multi with cohort-aware AUTO selection; Mapping for
    platform-agnostic preprocessing.
  - Interactive Shiny app with Bioconductor-friendly exports.

Upgrade Notes

  - Raw RNA-seq counts are supported from v1.1.3 onward (requires gene
    lengths).
  - If you previously parsed BS / BS.Subtype, switch to Call_5class /
    Call_4class.
  - Package API unchanged.